NewCloud data lakes, DuckDB SQL & dashboards

Explore Parquet
beautifully

The native desktop explorer for Parquet files and data lakes. Stream 50 GB files without loading them, filter and sort in Rust, query with SQL, browse S3 — your data never leaves your machine.

One payment · 3 devices · V1.x updates included

events_2026.parquet — 48.2 GB · 812,455,209 rows
Search all columns…amount > 1000city ↓scan 41 ms
id
name
city
amount
is_active
1042
user_1042
Paris
8,204.50
true
1043
user_1043
Douala
912.00
false
1044
user_1044
Abidjan
3,377.25
true
1045
user_1045
Dakar
156.80
true
1046
user_1046
Lomé
7,940.10
false
1047
user_1047
Bamako
2,051.90
true
500 loaded · 812,455,209 total s3://prod-lake connected

Works with

Apache ParquetDelta LakeIcebergDuckDBAmazon S3Cloudflare R2MinIOGCSAzure Blob

Built in Rust + Tauri

Everything a data engineer expects

Feels like Linear. Works like a database console. Weighs less than your browser tab.

Instant on huge files

Offset pushdown skips whole row groups: jumping to row 100,000,000 only decodes the page you look at. Memory stays flat, whatever the file size.

48.2 GB file — 12.4 MB read · 99.97% skipped

Local-first & private

Files never leave your machine. Cloud credentials live in the OS keychain. Zero telemetry.

Server-grade filtering

11 operators, multi-column sort, global search — executed in Rust on Arrow batches.

containsbetweenis nullin list

SQL with DuckDB

JOIN, GROUP BY and window functions across multiple Parquet files. Saved queries, history, notebooks.

SELECT city, sum(amount) AS revenue
FROM data GROUP BY city ORDER BY revenue DESC
-- 8 rows · 213 ms

Your whole data lake

Browse buckets, preview remote Parquet by fetching only the bytes you view, upload & download with a real transfer queue.

S3R2MinIOGCSAzure

Stats, charts & dashboards

Column distributions, top values, KPIs — computed by streaming passes that never materialize your dataset.

Streaming engine

Fast is a feature

Data is decoded batch by batch and never fully loaded — sorting keeps a bounded top-K window, filters run on Arrow buffers, and remote files come down as HTTP range requests. Real numbers, measured on the engine.

230 msto page row 2,900,000
2.7 sto sort 3,000,000 rows
0 bytesof your data uploaded
3devices per license, forever

From raw files to answers

Attach files as tables, write SQL, chart the result — without leaving the app or spinning up a cluster.

SQL — revenue_by_city8 rows · 213 ms
SELECT city, count(*) AS orders, sum(amount) AS revenue
FROM data WHERE is_active
GROUP BY city ORDER BY revenue DESC;
city
orders
revenue
Paris
128,441
1,204,882.50
Douala
97,210
918,204.00
Abidjan
84,002
877,391.25
Dakar
61,876
512,940.75

One price. Yours forever.

No subscription. No seat management. Buy it once like software used to work.

Parqlab PersonalLifetime
$59one payment, forever
  • Unlimited rows, files and file sizes
  • All cloud providers (S3, R2, MinIO, GCS, Azure)
  • DuckDB SQL editor, notebooks & dashboards
  • CSV / JSON / Excel / Parquet exports
  • Compare, merge, split & schema tools
  • Use on up to 3 devices
  • All V1.x updates included

Secure payment by Stripe · Key delivered by email in seconds · 14-day refund policy

Frequently asked questions

Stop fighting your Parquet files

Join the data engineers who open, query and ship their lake data without spinning up a cluster.

Try it free first